Abstract
During the gas turbine design process, control and optimization of turbulence characteristics at the combustor-turbine interface can lead to improved specific fuel consumption. Here, we use computational large-eddy simulations to explore the turbulent flow characteristics at the exit plane of two combustor designs: a baseline design corresponding to an idealized non-reacting combustor simulator and a design that has been optimized to minimize turbulence intensity at the combustor exit while maintaining good mixing uniformity and low pressure drop. By comparing the two designs, we demonstrate that the size, orientation, and positioning of the primary zone and dilution jets are the primary drivers determining turbulence characteristics at the combustor exit. We show that by modifying the combustor geometry we can shift the energy-containing eddies to larger scales, affecting both the turbulence decay rate and the interaction of turbulence with the turbine boundary layer. In the optimized design, turbulence decays at a faster rate towards isotropic flow and we observe that Reynolds stresses are substantially decreased. We introduce a flat plate with a tapered leading edge downstream of the combustor exit to quantify the subsequent boundary layer profile, skin friction, and boundary layer dissipation for each design. This work illustrates how understanding the mechanisms responsible for decreased turbulence at the combustor exit can aid designers in maximizing performance of gas turbines, as well as demonstrating the potential of computational optimization for combustor design.
